Disaster recovery site requirements: Building your own DR site
Take these physical, operational and functional concerns into consideration before you build your own recovery site.
What you will learn in this tip: Building your own disaster recovery (DR) site is an expensive endeavor, and there are a number of issues you need to consider. In this tip, you will learn about what disaster recovery site requirements to be aware of when building your own disaster recovery site.
Constructing your own disaster recovery site may seem appealing because of the ability to create a site specific to your needs, not having to deal with a vendor and having complete control over your DR staff. However, building your own DR site is a big step. Before you take on the task of planning and building your own DR site, you should take these physical, operational and functional costs into consideration first.
Maintaining power at your disaster recovery site
Power is an essential disaster recovery site requirement. Without power there is no data center. Without guaranteed power, the data center cannot function. These are two simple sentences yet the concept will cover a number of issues. First, you must determine the hardware configuration your operation requires and what supporting equipment is necessary. Determining the hardware is a relatively simple task. See what applications are needed and what supporting hardware will be required to maintain this level of operation. Your applications and hardware will most likely be a subset of your current data center.
 |
||||
|
|
|||
|
||||
Once you have determined the proper hardware configuration for your organization, you then need to estimate potential growth. Get an electrical professional in to assist in determining the amount of power drain you will use by adding up the known wattage of the hardware and add a cushion. This estimation of the load will also help you determine the size of the uninterruptible power supply (UPS) you will need. UPS systems are a good thing to have, but keep in mind that they just act as a standby battery which needs to maintain a charge. Therefore, UPS units large enough to provide sufficient power to a data center operation require routine maintenance. They should be checked for proper level of fluids, corrosive buildup and other everyday maintenance issues. The other major issue with UPS units is the estimate of time to drain. For example, if you require 100 watts you need more than 100 watts of UPS power if you expect the power to be provided for an extended time.
Since the UPS is expected to protect you from a power loss for a limited time, you need to look at a generator for additional power. Keep in mind that the size of the generator is dependent directly upon the equipment you want to provide power to. Aside from the power creation concerns of a generator, you need to decide if it will require manual intervention to start up (some companies have manual start for a generator since they expect UPS to keep power on), or if it will start automatically in a predetermined fashion. Obviously if manual intervention is required, someone must be on site to start the unit. Sufficiently sized generators that are designed to run a data center are usually diesel generators. If you want the generator to operate properly for an extended period of time, make sure that additional fuel can be delivered. This means that a delivery contact should be in place. Also, access to the generator's fuel storage tank should be reachable in all types of weather. Lastly, make sure that there are multiple feeds of electricity from your local utility. You should not be on the same electric grid as your primary site.
Heating and cooling your disaster recovery site
In the initial days of mainframe operations, data centers required substantial heating and cooling systems. The mainframes which defined the data center initially produced a fair amount of heat and were finicky about proper environmental controls such as heat and humidity. Today with the deployment of servers, the environment is somewhat less of a factor. Today's servers are much more tolerant of temperature range, however, the servers present a different temperature issue, the creation of hot and cold spots. The air in today's data centers needs to be circulated in a way that the cold and hot air can flow properly. I have been in a number of data centers where, depending on where you stood in the room, the temperature fluctuated several degrees from one side or corner to another. Make sure your heating, ventilation and air conditioning (HVAC) system properly circulates your air so that the temperature in your data center is level in all sections.
Raised floors and dropped ceilings in data centers
Data centers and disaster recovery sites consist of a lot of cabling and wiring that should be placed under floor boards (raised flooring) for better protection. For safety reasons, when the wiring is below the floor you must put some water detection monitor on the real floor. Also, some data centers have air vents installed under the floor. If this is the case in your disaster recovery site, make sure when you design the layout of the room that no equipment is placed on the areas with floor vents. And be sure put floor vents where there may be hot spots.
Choosing the correct location for your recovery site
The location of a disaster recovery site should be a substantial distance away from the primary site so that a single event, such as a natural disaster, does not affect both locations. The distance should be close enough for staff to get to; but far enough so that the same incident does not affect both sites. However, keep in mind that your staff needs to reach the DR site within predefined times. When choosing the location, do not assume the ride will be smooth with no traffic. Some activities require manual intervention, so a "dark" data center may still require someone physically at the location. You also may require additional security staff for the DR location. Be aware of potential flood zones, freight rail tracks, Interstate highways that transfer hazardous materials etc, that could affect your location. It is wise to perform a risk assessment of the recovery site location as if it is your primary site.
Functional concerns
Even after taking physical and operational concerns into consideration, how do you know if it's wise to build your own disaster recovery site? Here is a list of additional questions you should ask yourself before building your own recovery site:
1. Do you need to maintain two sites? If so, is the location far enough so as not to be impacted by the same outage, yet close enough for staff to get to and set up within the given timeframe?
2. Is there true redundancy in the DR site?
3. Do the application time frames require maintaining your own disaster recovery site?
4. How does the financial impact differ from other alternatives?
Keep in mind that you are in control of the staff, and the location is yours, so you won't be dependent on a vendor. The main advantage of having your own site is that you can control all aspects of the disaster recovery process. And, if the above considerations work out in your favor, then build your own DR site.
About this author: Harvey Betan is a certified business continuity planning consultant with experience in disaster recovery in both technology and business functions. He migrated to BC after the restoration of a large insurance company with a major presence in the World Trade Center on Sept. 11. His career has spanned a dozen years in business continuity after a 15-year career as a senior manager in information technology for the financial, insurance and nonprofit sectors.
